Research On Elbow Recognition And Magnetic Abrasive Grinding Of Point Cloud Surface Feature

As the core power system of aerospace equipment,aero-engine transmits the working medium through the pipeline system.Therefore,in order to meet the space use requirements of the engine,the pipeline often adopts elbows with adaptability and high space utilization.Due to the different bending angles of the elbow,it is often limited by the processing and forming process.Quality defects such as microcracks and wrinkles are formed at the bending of the inner wall of the pipe,which need to be ground.In this thesis,the elbow recognition and magnetic particle grinding test method of point cloud surface feature is proposed.The midpoint of the elbow is generated by point cloud data,and the trajectory of the magnetic particle grinding position of the elbow is planned to achieve the purpose of finishing the inner wall of the elbow.In the process of industrial robot grinding elbows,there are random gap fluctuations in the manual point collection method.The positioning of the middle line of the elbow is inaccurate,the execution efficiency is low,and the pose control is difficult.To solve this problem,this thesis uses a three-dimensional laser scanner to scan the surface of the elbow to form a point cloud data.The middle line of the elbow is extracted from the point cloud data,and the spatial coordinates are converted to generate the grinding pose of the industrial robot.At the same time,the working gap between the yoke and the elbow is adjusted to optimize the magnetic particle grinding trajectory of the elbow and improve the work efficiency and machining accuracy.According to the fluid characteristics of the working medium in the elbow,the CFD simulation is carried out to analyze the velocity and pressure distribution of the fluid in the elbow,and the erosion wear area of the inner wall of the elbow is sought.During the magnetic particle grinding,the position and posture of the inner wall of the elbow are adjusted by controlling the grinding gap,and the influence of different magnetic pole arrangement on the grinding area is explored.The magnetic field simulation analyzes the change of magnetic induction intensity in the elbow,and controls the position and posture trajectory of magnetic yoke grinding to ensure the uniform grinding of the inner wall of the elbow.Based on the point cloud data of the elbow profile,the B-spline surface reconstruction is carried out,the reconstruction model is segmented,and the point cloud data at the elbow bend is obtained.The intercepted elbow bend is downsampled,and the point cloud coordinate system is established by principal component analysis.At the same time,it is used as the coordinate system of the grinding workpiece.The motion pose of the industrial robot is solved by pose calculation,and the yoke is driven to process the inner wall of the elbow bend repeatedly.The test results show that the point cloud surface type identifies the middle line of the elbow,the grinding pose trajectory is generated with high accuracy,the yoke grinding movement is stable,and the surface quality of the inner wall of the elbow is uniform after processing.